CN116155329A - User clustering and power distribution method of mMIMO-NOMA system based on meta-heuristic algorithm - Google Patents

User clustering and power distribution method of mMIMO-NOMA system based on meta-heuristic algorithm Download PDF

Info

Publication number
CN116155329A
CN116155329A CN202310436619.6A CN202310436619A CN116155329A CN 116155329 A CN116155329 A CN 116155329A CN 202310436619 A CN202310436619 A CN 202310436619A CN 116155329 A CN116155329 A CN 116155329A
Authority
CN
China
Prior art keywords
user
cluster
algorithm
meta
users
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202310436619.6A
Other languages
Chinese (zh)
Other versions
CN116155329B (en
Inventor
李旺旺
黄学军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing University of Posts and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Posts and Telecommunications filed Critical Nanjing University of Posts and Telecommunications
Priority to CN202310436619.6A priority Critical patent/CN116155329B/en
Publication of CN116155329A publication Critical patent/CN116155329A/en
Application granted granted Critical
Publication of CN116155329B publication Critical patent/CN116155329B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H04B7/0456—Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H04B7/0426—Power distribution
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H04B7/0452—Multi-user MIMO systems
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Power Engineering (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Supply And Distribution Of Alternating Current (AREA)
  • Transmitters (AREA)

Abstract

本发明公开了一种基于元启发算法的mMIMO‑NOMA系统的用户分簇和功率分配方法，该方法包括以下步骤：步骤一，构建毫米波mMIMO‑NOMA系统，构建毫米波信道模型；步骤二，采用基于簇头选择的用户分簇算法，对所有用户进行分簇，得到用户分簇结果；步骤三，针对获得的簇头信道进行混合预编码，消除簇间的用户干扰；步骤四，使用基于融合PSO‑SCSO的元启发算法进行功率分配，提高系统的频谱效率和能量效率。本发明适用于多用户毫米波mMIMO‑NOMA系统，可以有效提升系统的频谱效率和能量效率。

Figure 202310436619

The invention discloses a method for user clustering and power allocation of an mMIMO-NOMA system based on a meta-heuristic algorithm. The method includes the following steps: step 1, constructing a millimeter-wave mMIMO-NOMA system, and constructing a millimeter-wave channel model; step 2, Use the user clustering algorithm based on cluster head selection to cluster all users to obtain the user clustering results; step 3, perform hybrid precoding on the obtained cluster head channel to eliminate user interference between clusters; step 4, use The meta-heuristic algorithm of PSO‑SCSO is used for power allocation, which improves the spectrum efficiency and energy efficiency of the system. The present invention is applicable to a multi-user millimeter wave mMIMO-NOMA system, and can effectively improve the spectrum efficiency and energy efficiency of the system.

Figure 202310436619

Description

基于元启发算法的mMIMO-NOMA系统的用户分簇和功率分配 方法User clustering and power allocation for mMIMO-NOMA systems based on meta-heuristic algorithm Method

技术领域Technical Field

本发明属于毫米波通信技术领域，具体涉及一种基于元启发算法的mMIMO-NOMA系统的用户分簇和功率分配方法。The present invention belongs to the technical field of millimeter wave communications, and in particular relates to a user clustering and power allocation method for an mMIMO-NOMA system based on a meta-heuristic algorithm.

背景技术Background Art

目前，无线通信系统以正交多址接入方式为主，在此方式下，频谱效率较低，用户接入数受到限制。毫米波技术可以提供更加丰富的频谱资源；大规模多输入多输出(massive multiple input multiple output, mMIMO)技术利用空分复用提高频谱效率的同时也可以弥补毫米波的路径损耗；非正交多址接入(non-orthogonal multiple access，NOMA)技术通过串行干扰消除技术实现功率域复用，让多个用户共享同一时频资源，可以有效提升系统同时连接数。因此将毫米波MIMO与NOMA相结合，即毫米波mMIMO-NOMA系统，利用MIMO 的天线阵列，采用分簇方式实现 SDMA 和 NOMA 的混合多址，可以突破射频链的数目对用户连接数的限制，有望为未来无线网络提供更高速率和更低功耗的数据传输。At present, wireless communication systems are mainly based on orthogonal multiple access, under which the spectrum efficiency is low and the number of user access is limited. Millimeter wave technology can provide richer spectrum resources; massive multiple input multiple output (mMIMO) technology uses space division multiplexing to improve spectrum efficiency while also compensating for millimeter wave path loss; non-orthogonal multiple access (NOMA) technology uses serial interference cancellation technology to achieve power domain multiplexing, allowing multiple users to share the same time-frequency resources, which can effectively increase the number of simultaneous connections in the system. Therefore, combining millimeter wave MIMO with NOMA, that is, the millimeter wave mMIMO-NOMA system, uses MIMO antenna arrays and adopts clustering to achieve hybrid multiple access of SDMA and NOMA, which can break through the limitation of the number of RF chains on the number of user connections, and is expected to provide higher-speed and lower-power data transmission for future wireless networks.

在mMIMO-NOMA通信系统中，随着用户数的增加，用户间干扰会显著影响系统性能，不同簇间的干扰可以通过混合预编码技术解决，簇内用户间的干扰需要通过合理的用户分簇和功率分配算法解决。近年来，国内外学者针对混合预编码、用户分簇和功率分配做了大量的研究，其中更多的研究集中在混合预编码上，也有较多学者研究了用户分簇和功率分配，提出了多种方案，但现有功率分配问题主要通过凸优化方法解决，计算复杂度高，传统基于机器学习的用户分簇算法也需要较为复杂的计算；近年来有学者提出了使用元启发算法求解NOMA系统功率分配问题，但是在mMIMO-NOMA系统中，用户数增加，传统元启发算法本身存在的缺陷导致性能下降，因此为mMIMO-NOMA系统设计一种高效的用户分簇和功率分配算法具有重要的意义。In the mMIMO-NOMA communication system, as the number of users increases, the interference between users will significantly affect the system performance. The interference between different clusters can be solved by hybrid precoding technology, and the interference between users in a cluster needs to be solved by reasonable user clustering and power allocation algorithms. In recent years, domestic and foreign scholars have done a lot of research on hybrid precoding, user clustering and power allocation, among which more research focuses on hybrid precoding. Many scholars have also studied user clustering and power allocation and proposed a variety of solutions, but the existing power allocation problem is mainly solved by convex optimization methods, which has high computational complexity. The traditional user clustering algorithm based on machine learning also requires relatively complex calculations. In recent years, some scholars have proposed using meta-heuristic algorithms to solve the power allocation problem of NOMA systems. However, in the mMIMO-NOMA system, the number of users increases, and the defects of the traditional meta-heuristic algorithm itself lead to performance degradation. Therefore, it is of great significance to design an efficient user clustering and power allocation algorithm for the mMIMO-NOMA system.

发明内容Summary of the invention

本发明针对mMIMO-NOMA系统中用户分簇和功率分配求解复杂的问题，提供了一种基于元启发算法的mMIMO-NOMA系统的用户分簇和功率分配方法，具体包括一种改进的基于簇头选择的用户分簇算法和融合粒子群算法（particle swarm optimization, PSO）和沙猫算法（Sand Cat Swarm Optimization, SCSO）的改进元启发算法用于功率分配方案，目的是降低计算复杂度，提高系统频谱效率和能量效率。The present invention aims at solving complex problems of user clustering and power allocation in mMIMO-NOMA system, and provides a user clustering and power allocation method for mMIMO-NOMA system based on meta-heuristic algorithm, specifically including an improved user clustering algorithm based on cluster head selection and an improved meta-heuristic algorithm integrating particle swarm optimization (PSO) and Sand Cat Swarm Optimization (SCSO) for power allocation scheme, aiming to reduce computational complexity and improve system spectrum efficiency and energy efficiency.

为实现上述目的，本发明提供了一种基于元启发算法的mMIMO-NOMA系统的用户分簇和功率分配方法，包括以下步骤：To achieve the above object, the present invention provides a user clustering and power allocation method for an mMIMO-NOMA system based on a meta-heuristic algorithm, comprising the following steps:

步骤一，构建毫米波mMIMO-NOMA系统，构建毫米波信道模型；Step 1: Build a millimeter wave mMIMO-NOMA system and a millimeter wave channel model;

步骤二，采用基于簇头选择的用户分簇算法，对所有用户进行分簇，得到用户分簇结果；Step 2: cluster all users using a user clustering algorithm based on cluster head selection to obtain user clustering results.

步骤三，针对获得的簇头信道进行混合预编码，消除簇间的用户干扰；Step 3: hybrid precoding is performed on the obtained cluster head channel to eliminate user interference between clusters;

步骤四，步骤四，使用基于融合PSO-SCSO的元启发算法进行功率分配。Step 4: In step 4, a meta-heuristic algorithm based on fusion PSO-SCSO is used to allocate power.

作为本发明的进一步改进，步骤一中，所述毫米波mMIMO-NOMA系统包括数字预编码模块、模拟预编码模块和G个用户簇，第

Figure SMS_1
簇中包含用户
Figure SMS_2
个，用户数据流根据分组和功率分配叠加之后流入数字预编码模块，然后流入模拟预编码模块，最终发送到各个用户。As a further improvement of the present invention, in step 1, the millimeter wave mMIMO-NOMA system includes a digital precoding module, an analog precoding module and G user clusters,
Figure SMS_1
The cluster contains users
Figure SMS_2
After the user data stream is grouped and superimposed on the power allocation, it flows into the digital precoding module, then flows into the analog precoding module, and is finally sent to each user.

作为本发明的进一步改进，簇

Figure SMS_3
中第
Figure SMS_4
个用户接收到的信号为：As a further improvement of the present invention, the cluster
Figure SMS_3
Middle
Figure SMS_4
The signal received by each user is:

Figure SMS_5
Figure SMS_5

其中，

Figure SMS_10
表示簇
Figure SMS_15
中用户
Figure SMS_23
的发射信号，
Figure SMS_8
表示簇
Figure SMS_17
中用户
Figure SMS_11
的接收信号；
Figure SMS_16
，
Figure SMS_20
，
Figure SMS_26
表示簇
Figure SMS_9
中用户
Figure SMS_13
的发射功率，
Figure SMS_22
表示簇
Figure SMS_30
中用户
Figure SMS_24
的发射功率，
Figure SMS_28
表示簇
Figure SMS_29
中用户
Figure SMS_34
的发射功率，
Figure SMS_36
表示簇
Figure SMS_41
中用户
Figure SMS_6
的发射信号，
Figure SMS_12
表示簇
Figure SMS_18
中用户
Figure SMS_21
的发射信号，
Figure SMS_35
是簇
Figure SMS_42
中用户
Figure SMS_27
的高斯噪声矢量，且
Figure SMS_33
；
Figure SMS_32
是模拟预编码矩阵，
Figure SMS_38
是矩阵的共轭转置操作，
Figure SMS_39
就是
Figure SMS_43
的共轭转置；
Figure SMS_19
表示数字预编码矩阵中的第
Figure SMS_25
列，
Figure SMS_7
表示数字预编码矩阵中的第
Figure SMS_14
列，
Figure SMS_31
表示簇
Figure SMS_37
中用户
Figure SMS_40
的信道矢量，采用均匀平面阵列的毫米波信道模型。in,
Figure SMS_10
Representation Cluster
Figure SMS_15
Medium User
Figure SMS_23
The transmission signal,
Figure SMS_8
Representation Cluster
Figure SMS_17
Medium User
Figure SMS_11
The received signal;
Figure SMS_16
,
Figure SMS_20
,
Figure SMS_26
Representation Cluster
Figure SMS_9
Medium User
Figure SMS_13
The transmission power,
Figure SMS_22
Representation Cluster
Figure SMS_30
Medium User
Figure SMS_24
The transmission power,
Figure SMS_28
Representation Cluster
Figure SMS_29
Medium User
Figure SMS_34
The transmission power,
Figure SMS_36
Representation Cluster
Figure SMS_41
Medium User
Figure SMS_6
The transmission signal,
Figure SMS_12
Representation Cluster
Figure SMS_18
Medium User
Figure SMS_21
The transmission signal,
Figure SMS_35
It is a cluster
Figure SMS_42
Medium User
Figure SMS_27
A Gaussian noise vector of
Figure SMS_33
;
Figure SMS_32
is the analog precoding matrix,
Figure SMS_38
is the conjugate transpose operation of the matrix,
Figure SMS_39
that is
Figure SMS_43
The conjugate transpose of ;
Figure SMS_19
represents the first
Figure SMS_25
List,
Figure SMS_7
represents the first
Figure SMS_14
List,
Figure SMS_31
Representation Cluster
Figure SMS_37
Medium User
Figure SMS_40
The channel vector of the millimeter wave channel model using a uniform planar array.

作为本发明的进一步改进，步骤二中采用基于簇头选择的用户分簇算法，对所有用户进行自适应分簇，具体包括：As a further improvement of the present invention, in step 2, a user clustering algorithm based on cluster head selection is used to adaptively cluster all users, specifically including:

利用毫米波的方向性特点，将用户根据信道相关性进行分簇，同一簇内的用户使用同一模拟预编码，即从同一波束中获得波束增益；同一簇内用户信道的相关性高，不同簇用户信道的相关性低；簇头用户为每簇中的强用户；具体算法过程如下：Taking advantage of the directional characteristics of millimeter waves, users are clustered according to channel correlation. Users in the same cluster use the same simulated precoding, that is, they obtain beam gain from the same beam. The correlation of user channels in the same cluster is high, and the correlation of user channels in different clusters is low. The cluster head user is a strong user in each cluster. The specific algorithm process is as follows:

Step1.初始化：初始化用户信道增益向量

Figure SMS_46
，其中
Figure SMS_49
；
Figure SMS_52
是第
Figure SMS_45
个用户的信道矢量，
Figure SMS_47
，
Figure SMS_50
表示用户总数；簇头集合
Figure SMS_53
初始为空集；初始化阈值
Figure SMS_44
；设置每簇中用户最大数
Figure SMS_48
；
Figure SMS_51
；Step 1. Initialization: Initialize the user channel gain vector
Figure SMS_46
,in
Figure SMS_49
;
Figure SMS_52
It is
Figure SMS_45
The channel vector of each user is
Figure SMS_47
,
Figure SMS_50
Indicates the total number of users; cluster head set
Figure SMS_53
Initially an empty set; initial threshold
Figure SMS_44
; Set the maximum number of users in each cluster
Figure SMS_48
;
Figure SMS_51
;

Step2.选择当前信道增益向量中最大元素对应的信道

Figure SMS_54
作为当前簇头，并将其从信道集合和信道增益向量中去除；Step 2. Select the channel corresponding to the largest element in the current channel gain vector
Figure SMS_54
As the current cluster head, and remove it from the channel set and channel gain vector;

Step3.计算信道集合中剩余所有用户信道

Figure SMS_55
与当前簇头的相关性
Figure SMS_56
，当且仅当该簇中用户数不超过
Figure SMS_57
并且
Figure SMS_58
时，将
Figure SMS_59
对应的用户与当前簇头对应用户归入第
Figure SMS_60
簇，并将其从剩余用户信道集合中去除；Step 3. Calculate all remaining user channels in the channel set
Figure SMS_55
Correlation with the current cluster head
Figure SMS_56
, if and only if the number of users in the cluster does not exceed
Figure SMS_57
and
Figure SMS_58
When
Figure SMS_59
The corresponding user and the current cluster head corresponding user are classified into the
Figure SMS_60
cluster and remove it from the remaining user channel set;

Step4.

Figure SMS_61
；Step4.
Figure SMS_61
;

Step5.重复Step3和Step4，直到所有用户都已经完成分簇，分簇结束，设所有用户一共被分为

Figure SMS_62
簇，第
Figure SMS_63
簇中包含用户
Figure SMS_64
个，则分簇后所有用户用
Figure SMS_65
表示。Step 5. Repeat Step 3 and Step 4 until all users have completed clustering. The clustering is completed. Suppose all users are divided into
Figure SMS_62
Cluster,
Figure SMS_63
The cluster contains users
Figure SMS_64
After clustering, all users
Figure SMS_65
express.

作为本发明的进一步改进，步骤三中使用混合预编码，包括模拟预编码和数字预编码，其中，所述模拟预编码使用移相器实现，仅调整信号的相位；所述数字预编码通过射频链实现，以同时调整相位和幅度。As a further improvement of the present invention, hybrid precoding is used in step three, including analog precoding and digital precoding, wherein the analog precoding is implemented using a phase shifter to only adjust the phase of the signal; and the digital precoding is implemented through a radio frequency chain to simultaneously adjust the phase and amplitude.

作为本发明的进一步改进，步骤四中以最大化系统频谱效率和能量效率为目标，采用融合PSO-SCSO的元启发算法求解用户功率分配，通过对粒子运动方式进行改进，并且融合SCSO算法，可以在更少次数的迭代之后获得更精确的结果。As a further improvement of the present invention, in step 4, with the goal of maximizing the system spectrum efficiency and energy efficiency, a meta-heuristic algorithm integrating PSO-SCSO is used to solve the user power allocation. By improving the particle motion mode and integrating the SCSO algorithm, more accurate results can be obtained after fewer iterations.

作为本发明的进一步改进，所述融合PSO-SCSO的元启发算法包括：As a further improvement of the present invention, the meta-heuristic algorithm of the fusion PSO-SCSO includes:

融合PSO-SCSO算法将粒子群算法PSO和沙猫优化算法SCSO相结合，利用SCSO的高维搜索能力提高PSO的开发能力和全局搜索能力；融合PSO-SCSO算法利用改进的方式更新粒子位置，其算法步骤如下：The fusion PSO-SCSO algorithm combines the particle swarm algorithm PSO and the sand cat optimization algorithm SCSO, and uses the high-dimensional search capability of SCSO to improve the development and global search capabilities of PSO; the fusion PSO-SCSO algorithm uses an improved method to update the particle position. The algorithm steps are as follows:

Step1.初始化粒子种群的大小，初始化所有的参数，随机初始化粒子群；Step 1. Initialize the size of the particle population, initialize all parameters, and randomly initialize the particle population;

Step2.计算所有粒子的适应度值，如果优于全局最优位置的适应度值，则更新全局最优位置；Step 2. Calculate the fitness value of all particles. If it is better than the fitness value of the global optimal position, update the global optimal position;

Step3.利用如下公式更新所有粒子的位置；Step 3. Update the positions of all particles using the following formula;

Figure SMS_66
Figure SMS_66

其中，

Figure SMS_81
表示第
Figure SMS_71
个粒子在第
Figure SMS_76
次迭代过程中的位置矢量；
Figure SMS_80
表示第
Figure SMS_85
个粒子在第
Figure SMS_84
次迭代过程中的位置矢量；
Figure SMS_86
为引入的一个矢量；
Figure SMS_72
、
Figure SMS_78
、
Figure SMS_67
都是0到1之间服从均匀分布的随机数，
Figure SMS_73
为0到
Figure SMS_70
之间服从均匀分布的随机值；
Figure SMS_77
、
Figure SMS_79
均是公式的中间变量，分别表示粒子在运动前期和后期的主要位置更新方式；
Figure SMS_83
是每次迭代过程中的全局最优位置矢量；
Figure SMS_69
为一个标量，初始值为
Figure SMS_74
，迭代过程中逐渐减小；
Figure SMS_75
是一个控制系数；
Figure SMS_82
和
Figure SMS_68
均为加速因子；in,
Figure SMS_81
Indicates
Figure SMS_71
The particle in
Figure SMS_76
The position vector during the iteration;
Figure SMS_80
Indicates
Figure SMS_85
The particle in
Figure SMS_84
The position vector during the iteration;
Figure SMS_86
is a vector introduced;
Figure SMS_72
,
Figure SMS_78
,
Figure SMS_67
They are all random numbers between 0 and 1 that follow a uniform distribution.
Figure SMS_73
0 to
Figure SMS_70
A random value that follows a uniform distribution between
Figure SMS_77
,
Figure SMS_79
They are all intermediate variables in the formula, representing the main position update methods of particles in the early and late stages of movement respectively;
Figure SMS_83
is the global optimal position vector in each iteration;
Figure SMS_69
is a scalar with an initial value of
Figure SMS_74
, gradually decreases during the iteration process;
Figure SMS_75
is a control coefficient;
Figure SMS_82
and
Figure SMS_68
All are acceleration factors;

Step4.重复Step2、Step3直到算法收敛；Step4. Repeat Step2 and Step3 until the algorithm converges;

Step5.输出算法更新位置信息。Step 5. Output algorithm to update position information.

本发明的有益效果为：本发明适用于毫米波mMIMO-NOMA多用户系统，采用基于簇头选择的用户分簇算法对用户分簇，以最大化频谱效率和能量效率加权和为目标，采用改进的元启发算法进行功率分配；所述元启发算法与传统元启发算法相比，表现出更精确的搜索结果和较快的搜索速度；其用于系统功率分配，可以使系统获得更高的频谱效率和能量效率，并且减少计算的复杂度。The beneficial effects of the present invention are as follows: the present invention is applicable to a millimeter wave mMIMO-NOMA multi-user system, and adopts a user clustering algorithm based on cluster head selection to cluster users, with the goal of maximizing the weighted sum of spectrum efficiency and energy efficiency, and adopts an improved meta-heuristic algorithm for power allocation; compared with traditional meta-heuristic algorithms, the meta-heuristic algorithm exhibits more accurate search results and faster search speed; when used for system power allocation, it can enable the system to obtain higher spectrum efficiency and energy efficiency, and reduce the complexity of calculation.

附图说明BRIEF DESCRIPTION OF THE DRAWINGS

图1是本发明实施例中的基于元启发算法的mMIMO-NOMA系统的用户分簇和功率分配方法的流程图。FIG1 is a flow chart of a method for user clustering and power allocation in an mMIMO-NOMA system based on a meta-heuristic algorithm in an embodiment of the present invention.

图2是本发明实施例中的毫米波mMIMO-NOMA系统模型图。FIG2 is a diagram of a millimeter wave mMIMO-NOMA system model in an embodiment of the present invention.

图3是本发明实施例中融合PSO-SCSO算法的元启发算法的算法流程图。FIG3 is an algorithm flow chart of a meta-heuristic algorithm integrating a PSO-SCSO algorithm in an embodiment of the present invention.

图4是本发明实施例中算法收敛性分析图。FIG. 4 is a graph showing the convergence analysis of the algorithm in the embodiment of the present invention.

图5是本发明实施例中所提出功率分配算法的系统频谱效率与信噪比关系对比示意图。FIG5 is a schematic diagram showing a comparison of the relationship between the system spectrum efficiency and the signal-to-noise ratio of the power allocation algorithm proposed in the embodiment of the present invention.

图6是本发明实施例中所提出功率分配算法的系统能量效率与信噪比关系对比示意图。FIG6 is a schematic diagram showing a comparison of the relationship between the system energy efficiency and the signal-to-noise ratio of the power allocation algorithm proposed in the embodiment of the present invention.

图7是本发明实施例中所提出用户分簇算法的系统频谱效率与信噪比关系对比示意图。FIG. 7 is a schematic diagram showing a comparison of the relationship between the system spectrum efficiency and the signal-to-noise ratio of the user clustering algorithm proposed in the embodiment of the present invention.

图8是本发明实施例中所提出用户分簇算法的系统能量效率与信噪比关系对比示意图。FIG8 is a schematic diagram showing a comparison of the relationship between the system energy efficiency and the signal-to-noise ratio of the user clustering algorithm proposed in the embodiment of the present invention.

具体实施方式DETAILED DESCRIPTION

为了使本发明的目的、技术方案和优点更加清楚，下面结合附图和具体实施例对本发明进行详细描述。In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

如图1所示，本发明提供了一种基于元启发算法的mMIMO-NOMA系统的用户分簇和功率分配方法，主要包括以下步骤：As shown in FIG1 , the present invention provides a user clustering and power allocation method for an mMIMO-NOMA system based on a meta-heuristic algorithm, which mainly includes the following steps:

步骤一，构建毫米波mMIMO-NOMA系统，构建毫米波信道模型；Step 1: Build a millimeter wave mMIMO-NOMA system and a millimeter wave channel model;

步骤二，采用基于簇头选择的用户分簇算法，对所有用户进行分簇，得到用户分簇结果；Step 2: cluster all users using a user clustering algorithm based on cluster head selection to obtain user clustering results.

步骤三，针对获得的簇头信道进行混合预编码，消除簇间的用户干扰；Step 3: hybrid precoding is performed on the obtained cluster head channel to eliminate user interference between clusters;

步骤四，使用基于融合PSO-SCSO的元启发算法进行功率分配。Step 4: Use the meta-heuristic algorithm based on the fusion PSO-SCSO to allocate power.

以下将结合附图对步骤一~步骤四进行详细描述。The following will describe steps 1 to 4 in detail with reference to the accompanying drawings.

步骤一中，所述毫米波mMIMO-NOMA系统包括数字预编码模块、模拟预编码模块和G个用户簇，第

Figure SMS_87
簇中包含用户
Figure SMS_88
个，用户数据流根据分组和功率分配叠加之后流入数字预编码模块，然后流入模拟预编码模块，最终发送到各个用户。In step 1, the millimeter wave mMIMO-NOMA system includes a digital precoding module, an analog precoding module and G user clusters.
Figure SMS_87
The cluster contains users
Figure SMS_88
After the user data stream is grouped and superimposed on the power allocation, it flows into the digital precoding module, then flows into the analog precoding module, and is finally sent to each user.

也就是说，步骤一具体为：构建如图2所示的多用户毫米波mMIMO-NOMA系统模型，BS端配有

Figure SMS_89
根发射天线和
Figure SMS_90
个RF链，同时服务
Figure SMS_91
个随机分布的单天线用户，
Figure SMS_92
Figure SMS_93
Figure SMS_94
。That is to say, step 1 is as follows: construct a multi-user millimeter wave mMIMO-NOMA system model as shown in Figure 2, with the BS end equipped with
Figure SMS_89
Root transmitting antenna and
Figure SMS_90
RF chains, serving
Figure SMS_91
randomly distributed single-antenna users,
Figure SMS_92
Figure SMS_93
Figure SMS_94
.

为了充分获得多路复用增益，设RF链的数量等于波束数量

Figure SMS_95
。To fully exploit the multiplexing gain, set the number of RF chains equal to the number of beams.
Figure SMS_95
.

通过NOMA技术，将用户分为

Figure SMS_96
簇，第
Figure SMS_97
簇中共计
Figure SMS_98
个用户共用同一波束。Through NOMA technology, users are divided into
Figure SMS_96
Cluster,
Figure SMS_97
Total in cluster
Figure SMS_98
Users share the same beam.

令

Figure SMS_99
、
Figure SMS_100
分别表示混合预编码中的模拟预编码矩阵和数字预编码矩阵，则簇
Figure SMS_101
中第
Figure SMS_102
个用户接收到的信号表示为：make
Figure SMS_99
,
Figure SMS_100
denote the analog precoding matrix and the digital precoding matrix in the hybrid precoding, respectively. Then the cluster
Figure SMS_101
Middle
Figure SMS_102
The signal received by a user is expressed as:

Figure SMS_103
（1）
Figure SMS_103
(1)

其中，

Figure SMS_132
表示簇
Figure SMS_137
中用户
Figure SMS_139
的发射信号，
Figure SMS_108
表示簇
Figure SMS_114
中用户
Figure SMS_120
的接收信号；
Figure SMS_127
，
Figure SMS_128
，
Figure SMS_133
表示簇
Figure SMS_135
中用户
Figure SMS_140
的发射功率，
Figure SMS_130
表示簇
Figure SMS_134
中用户
Figure SMS_138
的发射功率，
Figure SMS_142
表示簇
Figure SMS_111
中用户
Figure SMS_119
的发射功率，
Figure SMS_123
表示簇
Figure SMS_129
中用户
Figure SMS_104
的发射信号，
Figure SMS_110
表示簇
Figure SMS_116
中用户
Figure SMS_122
的发射信号，
Figure SMS_109
和
Figure SMS_115
的取值范围如累和符号中描述，
Figure SMS_117
是簇
Figure SMS_126
中用户
Figure SMS_131
的高斯噪声矢量，且
Figure SMS_136
；
Figure SMS_141
是模拟预编码矩阵，
Figure SMS_143
是矩阵的共轭转置操作，
Figure SMS_106
就是
Figure SMS_112
的共轭转置；
Figure SMS_121
表示数字预编码矩阵中的第
Figure SMS_125
列，
Figure SMS_107
表示数字预编码矩阵中的第
Figure SMS_113
列，
Figure SMS_118
表示簇
Figure SMS_124
中用户
Figure SMS_105
的信道矢量，采用均匀平面阵列的毫米波信道模型，则用户对应的信干噪比为：in,
Figure SMS_132
Representation Cluster
Figure SMS_137
Medium User
Figure SMS_139
The transmission signal,
Figure SMS_108
Representation Cluster
Figure SMS_114
Medium User
Figure SMS_120
The received signal;
Figure SMS_127
,
Figure SMS_128
,
Figure SMS_133
Representation Cluster
Figure SMS_135
Medium User
Figure SMS_140
The transmission power,
Figure SMS_130
Representation Cluster
Figure SMS_134
Medium User
Figure SMS_138
The transmission power,
Figure SMS_142
Representation Cluster
Figure SMS_111
Medium User
Figure SMS_119
The transmission power,
Figure SMS_123
Representation Cluster
Figure SMS_129
Medium User
Figure SMS_104
The transmission signal,
Figure SMS_110
Representation Cluster
Figure SMS_116
Medium User
Figure SMS_122
The transmission signal,
Figure SMS_109
and
Figure SMS_115
The value range of is as described in the cumulative symbol.
Figure SMS_117
It is a cluster
Figure SMS_126
Medium User
Figure SMS_131
A Gaussian noise vector of
Figure SMS_136
;
Figure SMS_141
is the analog precoding matrix,
Figure SMS_143
is the conjugate transpose operation of the matrix,
Figure SMS_106
that is
Figure SMS_112
The conjugate transpose of ;
Figure SMS_121
represents the first
Figure SMS_125
List,
Figure SMS_107
represents the first
Figure SMS_113
List,
Figure SMS_118
Representation Cluster
Figure SMS_124
Medium User
Figure SMS_105
The channel vector is , and the millimeter wave channel model of uniform planar array is adopted. Then the corresponding signal to noise ratio of the user is:

Figure SMS_144
（2）
Figure SMS_144
(2)

其中：in:

Figure SMS_145
（3）
Figure SMS_145
(3)

其中，

Figure SMS_146
表示数字预编码矩阵中的第
Figure SMS_147
列。in,
Figure SMS_146
represents the first
Figure SMS_147
List.

步骤二中采用基于簇头选择的用户分簇算法，对所有用户进行自适应分簇，得到用户分簇结果，具体方法如下：In step 2, a user clustering algorithm based on cluster head selection is used to adaptively cluster all users to obtain user clustering results. The specific method is as follows:

利用毫米波的方向性特点，将用户根据信道相关性进行分簇，同一簇内的用户使用同一模拟预编码，即从同一波束中获得波束增益；同一簇内用户信道的相关性高，不同簇用户信道的相关性低；簇头用户为每簇中的强用户；具体算法过程如下：Taking advantage of the directional characteristics of millimeter waves, users are clustered according to channel correlation. Users in the same cluster use the same simulated precoding, that is, they obtain beam gain from the same beam. The correlation of user channels in the same cluster is high, and the correlation of user channels in different clusters is low. The cluster head user is a strong user in each cluster. The specific algorithm process is as follows:

Step1.初始化：初始化用户信道增益向量

Figure SMS_149
，其中
Figure SMS_153
；
Figure SMS_156
是第
Figure SMS_150
个用户的信道矢量，
Figure SMS_152
，
Figure SMS_154
表示用户总数；簇头集合
Figure SMS_157
初始为空集；初始化阈值
Figure SMS_148
；设置每簇中用户最大数
Figure SMS_151
；
Figure SMS_155
；Step 1. Initialization: Initialize the user channel gain vector
Figure SMS_149
,in
Figure SMS_153
;
Figure SMS_156
It is
Figure SMS_150
The channel vector of each user is
Figure SMS_152
,
Figure SMS_154
Indicates the total number of users; cluster head set
Figure SMS_157
Initially an empty set; initial threshold
Figure SMS_148
; Set the maximum number of users in each cluster
Figure SMS_151
;
Figure SMS_155
;

Step2.选择当前信道增益向量中最大元素对应的信道

Figure SMS_158
作为当前簇头，并将其从信道集合和信道增益向量中去除；Step 2. Select the channel corresponding to the largest element in the current channel gain vector
Figure SMS_158
As the current cluster head, and remove it from the channel set and channel gain vector;

Step3.计算信道集合中剩余所有用户信道

Figure SMS_159
与当前簇头的相关性
Figure SMS_160
，当且仅当该簇中用户数不超过
Figure SMS_161
并且
Figure SMS_162
时，将
Figure SMS_163
对应的用户与当前簇头对应用户归入第
Figure SMS_164
簇，并将其从剩余用户信道集合中去除；Step 3. Calculate all remaining user channels in the channel set
Figure SMS_159
Correlation with the current cluster head
Figure SMS_160
, if and only if the number of users in the cluster does not exceed
Figure SMS_161
and
Figure SMS_162
When
Figure SMS_163
The corresponding user and the current cluster head corresponding user are classified into the
Figure SMS_164
cluster and remove it from the remaining user channel set;

Step4.

Figure SMS_165
；Step4.
Figure SMS_165
;

Step5.重复Step3和Step4，直到所有用户都已经完成分簇，分簇结束，设所有用户一共被分为

Figure SMS_166
簇，第
Figure SMS_167
簇中包含用户
Figure SMS_168
个，则分簇后所有用户用
Figure SMS_169
表示。Step 5. Repeat Step 3 and Step 4 until all users have completed clustering. The clustering is completed. Suppose all users are divided into
Figure SMS_166
Cluster,
Figure SMS_167
The cluster contains users
Figure SMS_168
After clustering, all users
Figure SMS_169
express.

步骤三中使用混合预编码，包括模拟预编码和数字预编码，其中，所述模拟预编码使用移相器实现，仅调整信号的相位；所述数字预编码通过射频链实现，以同时调整相位和幅度。In step three, hybrid precoding is used, including analog precoding and digital precoding, wherein the analog precoding is implemented using a phase shifter to adjust only the phase of the signal; the digital precoding is implemented through a radio frequency chain to simultaneously adjust the phase and amplitude.

步骤三具体为：针对获得的簇头信道进行混合预编码，消除簇间的用户干扰，由于模拟预编码矩阵

Figure SMS_170
只能够调整信号的相位，故而考虑使用信道矩阵的共轭转置的相位设计模拟预编码，同时考虑到移相器的精度问题，假设为
Figure SMS_171
比特精度的移相器，则模拟预编码矩阵可以表示为：Step 3 is as follows: hybrid precoding is performed on the obtained cluster head channel to eliminate user interference between clusters.
Figure SMS_170
Only the phase of the signal can be adjusted, so the phase design of the conjugate transpose of the channel matrix is considered to simulate the precoding. At the same time, considering the accuracy of the phase shifter, it is assumed that
Figure SMS_171
bit-precision phase shifter, the analog precoding matrix can be expressed as:

Figure SMS_172
（4）
Figure SMS_172
(4)

其中，

Figure SMS_174
是G个用户簇的簇头信道，
Figure SMS_178
表示
Figure SMS_181
的第
Figure SMS_175
行第
Figure SMS_177
个元素，
Figure SMS_180
表示
Figure SMS_183
的第
Figure SMS_173
行第
Figure SMS_176
个元素，
Figure SMS_179
是中间变量，
Figure SMS_182
表示计算复数的相位角。在获得了模拟预编码之后，得到所有簇头用户的等效信道为in,
Figure SMS_174
is the cluster head channel of G user clusters,
Figure SMS_178
express
Figure SMS_181
No.
Figure SMS_175
Line
Figure SMS_177
elements,
Figure SMS_180
express
Figure SMS_183
No.
Figure SMS_173
Line
Figure SMS_176
elements,
Figure SMS_179
is an intermediate variable,
Figure SMS_182
Represents the phase angle of the complex number. After obtaining the simulated precoding, the equivalent channels of all cluster head users are obtained as

Figure SMS_184
（5）
Figure SMS_184
(5)

则数字预编码矩阵为：Then the digital precoding matrix is:

Figure SMS_185
（6）
Figure SMS_185
(6)

步骤四中以最大化系统频谱效率和能量效率为目标，采用融合PSO-SCSO的元启发算法求解用户功率分配，通过对粒子运动方式进行改进，并且融合SCSO算法，可以在更少次数的迭代之后获得更精确的结果。In step 4, with the goal of maximizing the system spectrum efficiency and energy efficiency, a meta-heuristic algorithm integrating PSO-SCSO is used to solve the user power allocation. By improving the particle motion mode and integrating the SCSO algorithm, more accurate results can be obtained after fewer iterations.

所述融合PSO-SCSO的元启发算法包括：The meta-heuristic algorithm of the fusion PSO-SCSO includes:

融合PSO-SCSO算法将粒子群算法PSO和沙猫优化算法SCSO相结合，利用SCSO的高维搜索能力提高PSO的开发能力和全局搜索能力；融合PSO-SCSO算法利用改进的方式更新粒子位置，其算法步骤如下：The fusion PSO-SCSO algorithm combines the particle swarm algorithm PSO and the sand cat optimization algorithm SCSO, and uses the high-dimensional search capability of SCSO to improve the development and global search capabilities of PSO; the fusion PSO-SCSO algorithm uses an improved method to update the particle position. The algorithm steps are as follows:

Step1.初始化粒子种群的大小，初始化所有的参数，随机初始化粒子群；Step 1. Initialize the size of the particle population, initialize all parameters, and randomly initialize the particle population;

Step2.计算所有粒子的适应度值，如果优于全局最优位置的适应度值，则更新全局最优位置；Step 2. Calculate the fitness value of all particles. If it is better than the fitness value of the global optimal position, update the global optimal position;

Step3.利用如下公式更新所有粒子的位置；Step 3. Update the positions of all particles using the following formula;

Figure SMS_186
Figure SMS_186

其中，

Figure SMS_194
表示第
Figure SMS_188
个粒子在第
Figure SMS_198
次迭代过程中的位置矢量；
Figure SMS_189
表示第
Figure SMS_196
个粒子在第
Figure SMS_200
次迭代过程中的位置矢量；
Figure SMS_204
为引入的一个矢量，其定义在公式（17）；
Figure SMS_199
、
Figure SMS_203
、
Figure SMS_187
都是0到1之间服从均匀分布的随机数，
Figure SMS_193
为0到
Figure SMS_201
之间服从均匀分布的随机值；
Figure SMS_206
、
Figure SMS_202
均是公式的中间变量，分别表示粒子在运动前期和后期的主要位置更新方式；
Figure SMS_205
是每次迭代过程中的全局最优位置矢量；
Figure SMS_191
为一个标量，初始值为
Figure SMS_195
，迭代过程中逐渐减小；
Figure SMS_192
是一个控制系数；
Figure SMS_197
和
Figure SMS_190
均为加速因子，其定义在公式（19）；in,
Figure SMS_194
Indicates
Figure SMS_188
The particle in
Figure SMS_198
The position vector during the iteration;
Figure SMS_189
Indicates
Figure SMS_196
The particle in
Figure SMS_200
The position vector during the iteration;
Figure SMS_204
is a vector introduced and defined in formula (17);
Figure SMS_199
,
Figure SMS_203
,
Figure SMS_187
They are all random numbers between 0 and 1 that follow a uniform distribution.
Figure SMS_193
0 to
Figure SMS_201
A random value that follows a uniform distribution between
Figure SMS_206
,
Figure SMS_202
They are all intermediate variables in the formula, representing the main position update methods of particles in the early and late stages of movement respectively;
Figure SMS_205
is the global optimal position vector in each iteration;
Figure SMS_191
is a scalar with an initial value of
Figure SMS_195
, gradually decreases during the iteration process;
Figure SMS_192
is a control coefficient;
Figure SMS_197
and
Figure SMS_190
are acceleration factors, which are defined in formula (19);

Step4.重复Step2、Step3直到算法收敛；Step4. Repeat Step2 and Step3 until the algorithm converges;

Step5.输出算法更新位置信息。Step 5. Output algorithm to update position information.

具体来说，步骤四中使用基于融合PSO-SCSO的元启发算法进行功率分配，以提高系统的频谱效率和能量效率。Specifically, in step 4, a meta-heuristic algorithm based on fusion PSO-SCSO is used for power allocation to improve the spectrum efficiency and energy efficiency of the system.

首先确定优化目标，在完成混合预编码之后，先对簇中的用户按信道增益进行排序并重新编号，排序之后的结果满足：First, determine the optimization goal. After completing hybrid precoding, sort the users in the cluster by channel gain and renumber them. The sorted results satisfy:

Figure SMS_207
Figure SMS_207

第g簇中第m个用户的信息传输速率表示如下：The information transmission rate of the mth user in the gth cluster is expressed as follows:

Figure SMS_208
（7）
Figure SMS_208
(7)

则系统的频谱效率表示为；Then the spectral efficiency of the system is expressed as;

Figure SMS_209
（8）
Figure SMS_209
(8)

系统的能量效率定义为每焦耳能量传输的比特数量，表达式如下：The energy efficiency of a system is defined as the number of bits transmitted per joule of energy, expressed as follows:

Figure SMS_210
（9）
Figure SMS_210
(9)

其中

Figure SMS_211
、
Figure SMS_212
、
Figure SMS_213
分别表示每个射频链功率、每个移相器功率和基带功率，
Figure SMS_214
表示移相器的数量。由于频谱效率和能量效率都是移动通信的关键指标，故本发明考虑以最大化它们的加权和作为优化目标，构建出如下优化问题：in
Figure SMS_211
,
Figure SMS_212
,
Figure SMS_213
Represents each RF chain power, each phase shifter power and baseband power respectively,
Figure SMS_214
Represents the number of phase shifters. Since spectrum efficiency and energy efficiency are both key indicators of mobile communications, the present invention considers maximizing their weighted sum as the optimization goal and constructs the following optimization problem:

Figure SMS_215
（10）
Figure SMS_215
(10)

其中，

Figure SMS_216
表示每个用户的发送功率应当为正数，
Figure SMS_217
表示所有用户的总发射功率小于基站最大发送功率
Figure SMS_218
，
Figure SMS_219
是第g簇中第m个用户的信息传输速率，见公式（7），
Figure SMS_220
保证每个用户的信息传输速率满足最低速率要求
Figure SMS_221
。in,
Figure SMS_216
Indicates that the transmit power of each user should be a positive number.
Figure SMS_217
Indicates that the total transmission power of all users is less than the maximum transmission power of the base station
Figure SMS_218
,
Figure SMS_219
is the information transmission rate of the mth user in the gth cluster, see formula (7),
Figure SMS_220
Ensure that each user's information transmission rate meets the minimum rate requirement
Figure SMS_221
.

为了方便求解，根据算法特点，忽略

Figure SMS_222
约束，利用罚函数将上述有约束最大化优化问题转化为无约束最小化优化问题：In order to facilitate the solution, according to the characteristics of the algorithm, ignore
Figure SMS_222
Constraints, using penalty functions to transform the above constrained maximization optimization problem into an unconstrained minimization optimization problem:

Figure SMS_223
（11）
Figure SMS_223
(11)

其中，

Figure SMS_225
表示系统频谱效率，见公式(8)；
Figure SMS_228
表示系统的能量效率，见公式(9)；ρ是惩罚因子；系统分为G簇，第
Figure SMS_230
簇中有
Figure SMS_226
个用户，
Figure SMS_227
是第g簇中第m个用户的发送功率，
Figure SMS_229
是系统总的发送功率约束；
Figure SMS_231
是第g簇中第m个用户的频谱效率，
Figure SMS_224
是满足各个用户要求的的最低频谱效率。in,
Figure SMS_225
represents the system spectrum efficiency, see formula (8);
Figure SMS_228
represents the energy efficiency of the system, see formula (9); ρ is the penalty factor; the system is divided into G clusters,
Figure SMS_230
In the cluster
Figure SMS_226
Users,
Figure SMS_227
is the transmit power of the mth user in the gth cluster,
Figure SMS_229
is the total transmit power constraint of the system;
Figure SMS_231
is the spectral efficiency of the mth user in the gth cluster,
Figure SMS_224
It is the minimum spectrum efficiency that meets the requirements of each user.

针对最小化优化问题（11），传统基于经典数学理论的优化算法计算过程复杂；而元启发算法通过全局随机搜索，可以通过简单的计算获得全局最优值，为了充分发挥算法的全局搜索能力，提高系统性能，改进PSO算法并且融合SCSO算法。For minimization optimization problems (11), the traditional optimization algorithm based on classical mathematical theory has a complex calculation process; while the meta-heuristic algorithm can obtain the global optimal value through simple calculations through global random search. In order to give full play to the global search capability of the algorithm and improve system performance, the PSO algorithm is improved and the SCSO algorithm is integrated.

PSO算法：PSO algorithm:

PSO算法从随机的初始值开始，通过追踪每次迭代过程中的局部最优解，最终确定全局最优解。其特点是结构简单、计算速度快，非常适合用于求解多目标优化问题。标准PSO算法中，令

Figure SMS_232
和
Figure SMS_233
分别表示第
Figure SMS_234
个粒子在第
Figure SMS_235
次迭代过程中的位置矢量和速度矢量，则第
Figure SMS_236
个粒子从第
Figure SMS_237
次迭代到第
Figure SMS_238
次的状态更新公式如下：The PSO algorithm starts with a random initial value and eventually determines the global optimal solution by tracking the local optimal solution in each iteration. It is characterized by simple structure and fast calculation speed, and is very suitable for solving multi-objective optimization problems. In the standard PSO algorithm, let
Figure SMS_232
and
Figure SMS_233
Respectively represent
Figure SMS_234
The particle in
Figure SMS_235
The position vector and velocity vector in the iteration process are
Figure SMS_236
Particles from
Figure SMS_237
Iteration to
Figure SMS_238
The state update formula is as follows:

Figure SMS_239
（12）
Figure SMS_239
(12)

其中，

Figure SMS_241
、
Figure SMS_244
是0到1之间服从均匀分布的随机数，
Figure SMS_247
表示第
Figure SMS_242
个粒子在第
Figure SMS_243
次迭代的最优位置，
Figure SMS_246
表示全局最优位置，
Figure SMS_248
为惯性权重，表示了对粒子此前运动状态的信任；
Figure SMS_240
，
Figure SMS_245
为加速因子，分别表示粒子对自身的经验与全局共享信息的信任。虽然PSO算法实现简单，收敛速度快，但它也有易陷入局部最优的缺点，这是因为PSO算法的粒子运动方向相对固定，使其易于早熟收敛。in,
Figure SMS_241
,
Figure SMS_244
is a random number between 0 and 1 that follows a uniform distribution.
Figure SMS_247
Indicates
Figure SMS_242
The particle in
Figure SMS_243
The optimal position of the iteration,
Figure SMS_246
represents the global optimal position,
Figure SMS_248
is the inertia weight, which represents the trust in the particle’s previous motion state;
Figure SMS_240
,
Figure SMS_245
are acceleration factors, and represent the particle's trust in its own experience and global shared information, respectively. Although the PSO algorithm is simple to implement and converges quickly, it also has the disadvantage of being easily trapped in local optimality. This is because the particle movement direction of the PSO algorithm is relatively fixed, making it prone to premature convergence.

SCSO算法：SCSO algorithm:

SCSO算法是2022年新提出的一种模仿沙猫生存行为的优化算法，具有收敛速度快、结果准确的特点，在高维和多目标优化问题中表现较好。令

Figure SMS_249
表示从第
Figure SMS_250
次迭代更新到第
Figure SMS_251
次迭代得到的种群的新位置，则SCSO算法的粒子更新公式如下所示。The SCSO algorithm is a new optimization algorithm proposed in 2022 that imitates the survival behavior of sand cats. It has the characteristics of fast convergence speed and accurate results, and performs well in high-dimensional and multi-objective optimization problems.
Figure SMS_249
Indicates that from
Figure SMS_250
Update to the next iteration
Figure SMS_251
The new position of the population is obtained by the iteration, and the particle update formula of the SCSO algorithm is as follows.

Figure SMS_252
（13）
Figure SMS_252
(13)

其中，

Figure SMS_254
为第
Figure SMS_256
次的全局最优解，
Figure SMS_259
为种群中成员
Figure SMS_255
时刻所处的位置，
Figure SMS_257
表示
Figure SMS_260
时刻各成员局部最优位置，
Figure SMS_261
是一个随机角度，用于控制群体中的每个成员在搜索空间中沿着不同的方向移动，
Figure SMS_253
和
Figure SMS_258
是0到1间的随机数。其他参数通过式(14~16)得到。in,
Figure SMS_254
For the
Figure SMS_256
The global optimal solution of
Figure SMS_259
For members of the population
Figure SMS_255
The location at the moment,
Figure SMS_257
express
Figure SMS_260
The local optimal position of each member at the moment,
Figure SMS_261
is a random angle used to control each member of the group to move in different directions in the search space.
Figure SMS_253
and
Figure SMS_258
is a random number between 0 and 1. Other parameters are obtained through equations (14~16).

Figure SMS_262
(14)
Figure SMS_262
(14)

Figure SMS_263
(15)
Figure SMS_263
(15)

Figure SMS_264
(16)
Figure SMS_264
(16)

其中，

Figure SMS_265
、
Figure SMS_266
是0到1之间的随机数，
Figure SMS_267
代表每只沙猫的敏感范围，一般设为2；
Figure SMS_268
、
Figure SMS_269
分别表示当前迭代次数和最大迭代次数，
Figure SMS_270
是中间变量，
Figure SMS_271
是用于控制沙猫行为的距离参数。in,
Figure SMS_265
,
Figure SMS_266
is a random number between 0 and 1,
Figure SMS_267
Represents the sensitivity range of each sand cat, usually set to 2;
Figure SMS_268
,
Figure SMS_269
Represent the current number of iterations and the maximum number of iterations respectively.
Figure SMS_270
is an intermediate variable,
Figure SMS_271
is the distance parameter used to control the behavior of the sand cat.

融合PSO-SCSO算法Fusion PSO-SCSO algorithm

根据以上描述，首先改进PSO算法，目的是加快算法的收敛速度和改善算法的全局搜索能力，改进其位置更新公式如下：According to the above description, the PSO algorithm is first improved to speed up the convergence speed of the algorithm and improve the global search ability of the algorithm. The position update formula is improved as follows:

Figure SMS_272
(17)
Figure SMS_272
(17)

其中，

Figure SMS_277
表示第
Figure SMS_274
个粒子在第
Figure SMS_279
次迭代过程的位置矢量，
Figure SMS_276
表示第
Figure SMS_281
个粒子在第
Figure SMS_286
次迭代过程中的位置矢量，
Figure SMS_290
表示所有粒子位置坐标的上边界，
Figure SMS_285
表示所有粒子位置坐标的下边界，
Figure SMS_289
表示全局最优解矢量，
Figure SMS_273
、
Figure SMS_280
、
Figure SMS_278
都是0到1之间的随机数，
Figure SMS_282
中元素均为0到
Figure SMS_283
之间的随机值，
Figure SMS_288
表示运动步长，
Figure SMS_275
控制收敛速度，
Figure SMS_284
初值为
Figure SMS_287
。in,
Figure SMS_277
Indicates
Figure SMS_274
The particle in
Figure SMS_279
The position vector of the iteration process,
Figure SMS_276
Indicates
Figure SMS_281
The particle in
Figure SMS_286
The position vector in the iteration process is
Figure SMS_290
represents the upper boundary of all particle position coordinates,
Figure SMS_285
represents the lower boundary of all particle position coordinates,
Figure SMS_289
represents the global optimal solution vector,
Figure SMS_273
,
Figure SMS_280
,
Figure SMS_278
are all random numbers between 0 and 1.
Figure SMS_282
The elements are all between 0 and
Figure SMS_283
A random value between
Figure SMS_288
represents the movement step length,
Figure SMS_275
Control the convergence speed,
Figure SMS_284
The initial value is
Figure SMS_287
.

式（17）中，第一项代替了原式中的惯性和局部最优因子，

Figure SMS_291
是指向全局最优解的向量，使用正弦函数作为系数，其结果促使粒子向着全局最优位置靠近或者远离，两者发生的概率比例为2：1，这样的设计加速了算法的收敛速度；第二项通过对粒子当前位置添加余弦扰动，其意义是使粒子从当前位置出发，随机向最优位置附近的范围内运动，使算法有更好的搜索能力；第三项中将原式中加速因子
Figure SMS_292
替换为一个正弦表达式，其值随着迭代次数增加而降低，使粒子在迭代初期快速靠近最优解，后期在全局最优点的附近缓慢收敛，从而避免了粒子在最优点附近震荡，改善了收敛性。In formula (17), the first term replaces the inertia and local optimal factor in the original formula,
Figure SMS_291
is a vector pointing to the global optimal solution, using the sine function as a coefficient. The result causes the particle to approach or move away from the global optimal position, with a probability ratio of 2:1. This design accelerates the convergence speed of the algorithm. The second term adds a cosine perturbation to the current position of the particle, which means that the particle starts from the current position and moves randomly within the range near the optimal position, giving the algorithm better search capabilities. The third term adds the acceleration factor in the original formula
Figure SMS_292
It is replaced by a sinusoidal expression, whose value decreases with the increase of iteration number, so that the particle quickly approaches the optimal solution in the early stage of iteration and slowly converges near the global optimal point in the later stage, thus avoiding the oscillation of particles near the optimal point and improving the convergence.

再参考SCSO算法中的攻击行为，修改式（17）中第三项，并且引入可变系数，再次改进之后的公式如下：Referring to the attack behavior in the SCSO algorithm, the third term in equation (17) is modified, and a variable coefficient is introduced. The improved formula is as follows:

Figure SMS_293
(18)
Figure SMS_293
(18)

其中，

Figure SMS_306
表示第
Figure SMS_297
个粒子在第
Figure SMS_300
次迭代过程中的位置矢量；
Figure SMS_309
表示第
Figure SMS_313
个粒子在第
Figure SMS_310
次迭代过程中的位置矢量；
Figure SMS_314
为引入的一个矢量，其定义与公式（17）中相同；
Figure SMS_307
、
Figure SMS_311
、
Figure SMS_294
都是0到1之间服从均匀分布的随机数，
Figure SMS_305
为0到
Figure SMS_298
之间服从均匀分布的随机值；
Figure SMS_301
、
Figure SMS_299
均是公式的中间变量，分别表示粒子在运动前期和后期的主要位置更新方式；
Figure SMS_304
是每次迭代过程中的全局最优位置矢量；
Figure SMS_296
为一个标量，初始值为
Figure SMS_302
，迭代过程中逐渐减小；
Figure SMS_308
是一个控制系数；
Figure SMS_312
和
Figure SMS_295
均为加速因子，值与
Figure SMS_303
相关，具体如下：in,
Figure SMS_306
Indicates
Figure SMS_297
The particle in
Figure SMS_300
The position vector during the iteration;
Figure SMS_309
Indicates
Figure SMS_313
The particle in
Figure SMS_310
The position vector during the iteration;
Figure SMS_314
is a vector introduced, and its definition is the same as that in formula (17);
Figure SMS_307
,
Figure SMS_311
,
Figure SMS_294
They are all random numbers between 0 and 1 that follow a uniform distribution.
Figure SMS_305
0 to
Figure SMS_298
A random value that follows a uniform distribution between
Figure SMS_301
,
Figure SMS_299
They are all intermediate variables in the formula, representing the main position update methods of particles in the early and late stages of movement respectively;
Figure SMS_304
is the global optimal position vector in each iteration;
Figure SMS_296
is a scalar with an initial value of
Figure SMS_302
, gradually decreases during the iteration process;
Figure SMS_308
is a control coefficient;
Figure SMS_312
and
Figure SMS_295
are acceleration factors, and their values are
Figure SMS_303
Related, as follows:

Figure SMS_315
(19)
Figure SMS_315
(19)

式（18）到（19）中，

Figure SMS_318
、
Figure SMS_321
、
Figure SMS_323
的值根据实际问题调整，
Figure SMS_317
的计算与沙猫算法中相同。改进后的算法以与全局最优点的距离为参数，若
Figure SMS_320
，
Figure SMS_322
，
Figure SMS_324
发挥主要作用，促使粒子向最优点靠近；否则
Figure SMS_316
,
Figure SMS_319
发挥主要作用，促使粒子在全局范围内搜索。In formulas (18) to (19),
Figure SMS_318
,
Figure SMS_321
,
Figure SMS_323
The value of is adjusted according to the actual problem.
Figure SMS_317
The calculation of is the same as in the Sand Cat algorithm. The improved algorithm uses the distance from the global optimal point as a parameter.
Figure SMS_320
,
Figure SMS_322
,
Figure SMS_324
Play a major role in driving particles closer to the optimal point; otherwise
Figure SMS_316
,
Figure SMS_319
Plays a major role in prompting particles to search globally.

所述融合PSO-SCSO算法的计算流程如图3所示。The calculation process of the fusion PSO-SCSO algorithm is shown in FIG3 .

以所有用户的发射功率作为算法中粒子的位置矢量，在经过有限次的迭代之后，算法收敛，输出结果即是用户的功率分配方案，可以最大化系统的频谱效率和能量效率。The transmission power of all users is used as the position vector of the particles in the algorithm. After a finite number of iterations, the algorithm converges and the output result is the user's power allocation plan, which can maximize the system's spectral efficiency and energy efficiency.

在上述实施例步骤下，通过在MATLAB平台进行仿真实验，从而说明本发明的有益效果。In the above-mentioned embodiment steps, simulation experiments are carried out on the MATLAB platform to illustrate the beneficial effects of the present invention.

下表展示了仿真参数设置，除表中系统参数，融合PSO-SCSO算法参数为：

Figure SMS_326
，
Figure SMS_328
，
Figure SMS_332
，
Figure SMS_327
；惩罚因子
Figure SMS_330
，
Figure SMS_331
。分簇算法中阈值
Figure SMS_333
的计算方式为随机选择
Figure SMS_325
个用户，令
Figure SMS_329
为这些用户中任意两用户间信道相关性的平均值的1.25倍，其中1.25为多次实验的经验值。The following table shows the simulation parameter settings. In addition to the system parameters in the table, the fusion PSO-SCSO algorithm parameters are:
Figure SMS_326
,
Figure SMS_328
,
Figure SMS_332
,
Figure SMS_327
; Penalty factor
Figure SMS_330
,
Figure SMS_331
. Threshold in clustering algorithm
Figure SMS_333
The calculation method is to randomly select
Figure SMS_325
Users, order
Figure SMS_329
It is 1.25 times the average value of the channel correlation between any two users among these users, where 1.25 is an empirical value obtained from multiple experiments.

Figure SMS_334
Figure SMS_334

图4为不同元启发算法的收敛性仿真图，对比了所提算法与经典元启发算法，包括PSO算法、灰狼优化（Grey Wolf Optimizer，GWO）算法，鲸鱼优化 (whale optimizationalgorithm, WOA)算法，从图中可以看出所提算法在大约10次以内即实现收敛，收敛速度最快且适应度值最低，验证了所提算法的收敛性。Figure 4 is a convergence simulation diagram of different meta-heuristic algorithms, which compares the proposed algorithm with classic meta-heuristic algorithms, including PSO algorithm, Grey Wolf Optimizer (GWO) algorithm, and whale optimization algorithm (WOA) algorithm. It can be seen from the figure that the proposed algorithm converges within about 10 times, with the fastest convergence speed and the lowest fitness value, which verifies the convergence of the proposed algorithm.

图5和图6分别为不同算法的能量效率和频谱效率与信噪比的关系，从图5中可以看出，全数字预编码的频谱效率最高，代表了理论上限，但是其成本高昂，难以应用于实际，所以仅供参考。在NOMA功率分配方案中，随着信噪比的增加，所提算法最接近全数字预编码，优于其他方案。图6展示了能量效率与信噪比的关系，从图6中可以看出，虽然全数字预编码的频谱效率最高，但是由于其需要大量的射频链来实现，所以其能量效率最低；而NOMA系统因为使用较少的射频链和利用了功率域复用，所以能量效率大大提高；且所提算法的能量效率要优于其他算法，这是因为在相同的功率消耗的情况下，所提算法能够获得更高的频谱效率。Figures 5 and 6 show the relationship between energy efficiency and spectrum efficiency and signal-to-noise ratio of different algorithms, respectively. As can be seen from Figure 5, the spectrum efficiency of full digital precoding is the highest, representing the theoretical upper limit, but its cost is high and difficult to apply in practice, so it is only for reference. In the NOMA power allocation scheme, as the signal-to-noise ratio increases, the proposed algorithm is closest to full digital precoding and is superior to other schemes. Figure 6 shows the relationship between energy efficiency and signal-to-noise ratio. As can be seen from Figure 6, although the spectrum efficiency of full digital precoding is the highest, it requires a large number of RF chains to implement, so its energy efficiency is the lowest; and the NOMA system uses fewer RF chains and utilizes power domain multiplexing, so the energy efficiency is greatly improved; and the energy efficiency of the proposed algorithm is better than other algorithms, because under the same power consumption, the proposed algorithm can obtain higher spectrum efficiency.

图7和图8比较了所提用户分簇算法与K均值分簇算法，K均值算法的簇数设为固定值6。从图7中可以看出，随着信噪比的增加，所提算法的频谱效率明显优于其他算法，这是因为本发明所提算法将相关性较高的用户信道分为一簇，否则单独作为一簇；而K均值算法将所有用户强行分为固定簇，导致存在簇内用户相关性低的情况，部分用户的传输速率低，也不利于簇内用户间干扰消除。图8中展示了不同分簇算法的能量效率，虽然所提算法的实际分簇数会高于其他算法，意味着需要更多的RF链和能量消耗，但从图中可以看出能量效率实际略高于K均值算法，所以这样的方案是合理的。FIG7 and FIG8 compare the proposed user clustering algorithm with the K-means clustering algorithm, and the number of clusters of the K-means algorithm is set to a fixed value of 6. As can be seen from FIG7, with the increase of the signal-to-noise ratio, the spectrum efficiency of the proposed algorithm is significantly better than that of other algorithms. This is because the algorithm proposed in the present invention divides the user channels with higher correlation into a cluster, otherwise they are treated as a single cluster; while the K-means algorithm forcibly divides all users into fixed clusters, resulting in low correlation among users within the cluster, low transmission rates for some users, and is not conducive to eliminating interference between users within the cluster. FIG8 shows the energy efficiency of different clustering algorithms. Although the actual number of clusters of the proposed algorithm will be higher than that of other algorithms, which means that more RF chains and energy consumption are required, it can be seen from the figure that the energy efficiency is actually slightly higher than that of the K-means algorithm, so such a solution is reasonable.

综上所述，本发明适用于毫米波mMIMO-NOMA多用户系统，采用基于簇头选择的用户分簇算法对用户分簇，以最大化频谱效率和能量效率加权和为目标，采用改进的元启发算法进行功率分配；所述元启发算法与传统元启发算法相比，表现出更精确的搜索结果和较快的搜索速度；其用于系统功率分配，可以使系统获得更高的频谱效率和能量效率，并且减少计算的复杂度。In summary, the present invention is suitable for millimeter wave mMIMO-NOMA multi-user systems, and adopts a user clustering algorithm based on cluster head selection to cluster users, with the goal of maximizing the weighted sum of spectral efficiency and energy efficiency, and adopts an improved meta-heuristic algorithm for power allocation; compared with the traditional meta-heuristic algorithm, the meta-heuristic algorithm shows more accurate search results and faster search speed; when used for system power allocation, it can enable the system to obtain higher spectral efficiency and energy efficiency and reduce the complexity of calculation.

以上实施例仅用以说明本发明的技术方案而非限制，尽管参照较佳实施例对本发明进行了详细说明，本领域的普通技术人员应当理解，可以对本发明的技术方案进行修改或者等同替换，而不脱离本发明技术方案的精神和范围。The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims (7)

1. The user clustering and power distribution method of the mMIMO-NOMA system based on the meta-heuristic algorithm is characterized by comprising the following steps:
firstly, constructing a millimeter wave mMIMO-NOMA system and constructing a millimeter wave channel model;
step two, clustering all users by adopting a user clustering algorithm based on cluster head selection to obtain a user clustering result;
step three, mixed pre-coding is carried out aiming at the obtained cluster head channels, and user interference among clusters is eliminated;
and step four, performing power distribution by using a meta-heuristic algorithm based on fusion PSO-SCSO.
2. The meta-heuristic based user clustering and power allocation method of a mimo-NOMA system of claim 1, wherein: in the first step, the millimeter wave mimo-NOMA system includes a digital precoding module, an analog precoding module, and G user clusters, the first step
Figure QLYQS_1
The cluster contains user->
Figure QLYQS_2
And the user data flow flows into a digital precoding module after being overlapped according to grouping and power distribution, then flows into an analog precoding module and finally is sent to each user.
3. The meta-heuristic based user clustering and power allocation method of a mimo-NOMA system of claim 2, whereinIn that, the cluster
Figure QLYQS_3
Middle->
Figure QLYQS_4
The signals received by the individual users are:
Figure QLYQS_9
wherein (1)>
Figure QLYQS_13
Representing cluster->
Figure QLYQS_19
Middle user->
Figure QLYQS_6
Is>
Figure QLYQS_12
Representing cluster->
Figure QLYQS_8
Middle user->
Figure QLYQS_15
Is a signal received by the base station;
Figure QLYQS_22
，
Figure QLYQS_28
，
Figure QLYQS_10
Representing cluster->
Figure QLYQS_16
Middle user->
Figure QLYQS_18
Transmit power of>
Figure QLYQS_23
Representing cluster->
Figure QLYQS_24
Middle user->
Figure QLYQS_31
Transmit power of>
Figure QLYQS_30
Representing cluster->
Figure QLYQS_38
Middle user->
Figure QLYQS_34
Transmit power of>
Figure QLYQS_40
Representing cluster->
Figure QLYQS_5
Middle user->
Figure QLYQS_11
Is>
Figure QLYQS_20
Representing cluster->
Figure QLYQS_26
Middle user->
Figure QLYQS_33
Is>
Figure QLYQS_39
Is cluster->
Figure QLYQS_21
Middle user->
Figure QLYQS_27
Is a Gaussian noise vector of>
Figure QLYQS_29
；
Figure QLYQS_35
Is an analog pre-coding matrix that is used to determine,
Figure QLYQS_36
is the conjugate transpose operation of the matrix,/->
Figure QLYQS_42
Namely +.>
Figure QLYQS_17
Is a conjugate transpose of (2);
Figure QLYQS_25
Representing the +.>
Figure QLYQS_7
Column (S)/(S)>
Figure QLYQS_14
Representing the +.>
Figure QLYQS_32
Column (S)/(S)>
Figure QLYQS_37
Representing cluster->
Figure QLYQS_41
Middle user->
Figure QLYQS_43
Adopts a uniform planeMillimeter wave channel model of the array.
4. The method for user clustering and power distribution of a mimo-NOMA system based on a meta-heuristic algorithm according to claim 1, wherein in the second step, a user clustering algorithm based on cluster head selection is adopted to perform adaptive clustering on all users, and specifically comprises:
clustering users according to channel correlation by utilizing the directivity characteristic of millimeter waves, wherein the users in the same cluster use the same analog precoding, namely, the beam gain is obtained from the same beam; the correlation of user channels in the same cluster is high, and the correlation of user channels in different clusters is low; the cluster head user is a strong user in each cluster; the specific algorithm process is as follows: step1. initializing: initializing user channel gain vectors
Figure QLYQS_45
Wherein->
Figure QLYQS_47
，
Figure QLYQS_50
Is->
Figure QLYQS_46
The channel vector of the individual user(s),
Figure QLYQS_49
，
Figure QLYQS_52
representing the total number of users; cluster head set->
Figure QLYQS_53
Initially empty; initialization threshold +.>
Figure QLYQS_44
The method comprises the steps of carrying out a first treatment on the surface of the Setting the maximum number of users in each cluster +.>
Figure QLYQS_48
；
Figure QLYQS_51
The method comprises the steps of carrying out a first treatment on the surface of the Step2, selecting the channel corresponding to the largest element in the current channel gain vector>
Figure QLYQS_54
As the current cluster head and removing it from the channel set and channel gain vector; />
Step3, calculating all remaining user channels in the channel set
Figure QLYQS_55
Correlation with current cluster head
Figure QLYQS_56
If and only if the number of users in the cluster does not exceed +.>
Figure QLYQS_57
And->
Figure QLYQS_58
When in use, will->
Figure QLYQS_59
The corresponding user is classified as the corresponding user of the current cluster head>
Figure QLYQS_60
Clusters and removing them from the remaining set of user channels; step4.
Figure QLYQS_61
；
Step5 repeating Step3 and Step4 until all users have completed clustering, and setting all users to be classified together
Figure QLYQS_62
Cluster, th->
Figure QLYQS_63
The cluster contains user->
Figure QLYQS_64
If yes, all users are used +.>
Figure QLYQS_65
And (3) representing.
5. The meta-heuristic based user clustering and power allocation method of a mimo-NOMA system of claim 1, wherein: step three, hybrid precoding is used, which comprises analog precoding and digital precoding, wherein the analog precoding is realized by using a phase shifter, and only the phase of a signal is adjusted; the digital precoding is implemented by a radio frequency chain to adjust both phase and amplitude.
6. The meta-heuristic based user clustering and power allocation method of a mimo-NOMA system of claim 1, wherein: in the fourth step, aiming at maximizing the spectral efficiency and the energy efficiency of the system, a meta-heuristic algorithm fused with PSO-SCSO is adopted to solve the user power distribution, and a particle motion mode is improved, and the SCSO algorithm is fused, so that a more accurate result can be obtained after fewer iterations.
7. The meta-heuristic method for user clustering and power allocation of a mimo-NOMA system based on the meta-heuristic algorithm of claim 6 wherein the meta-heuristic algorithm fusing PSO-SCSO comprises:
the PSO-SCSO algorithm is fused, the particle swarm algorithm PSO and the salsa optimization algorithm SCSO are combined, and the development capacity and the global searching capacity of the PSO are improved by utilizing the high-dimensional searching capacity of the SCSO; the fusion PSO-SCSO algorithm updates the particle position in an improved way, and comprises the following algorithm steps: step1, initializing the size of particle populations, initializing all parameters, and randomly initializing the particle populations;
step2, calculating the fitness value of all particles, and if the fitness value is better than the fitness value of the global optimal position, updating the global optimal position;
step3, updating the positions of all particles by using the following formula;
Figure QLYQS_80
wherein (1)>
Figure QLYQS_68
Indicate->
Figure QLYQS_74
The individual particles are at->
Figure QLYQS_71
Position vector in the iterative process of times;
Figure QLYQS_76
Indicate->
Figure QLYQS_81
The individual particles are at->
Figure QLYQS_85
Position vector in the iterative process of times;
Figure QLYQS_70
Is a vector introduced;
Figure QLYQS_77
、
Figure QLYQS_66
、
Figure QLYQS_75
Are random numbers between 0 and 1 subject to uniform distribution, ">
Figure QLYQS_69
From 0 to 0
Figure QLYQS_73
Random values subject to uniform distribution;
Figure QLYQS_79
、
Figure QLYQS_84
Are all intermediate variables of the formula, and respectively represent main position updating modes of the particles in the early stage and the later stage of movement;
Figure QLYQS_78
Is a global optimal position vector in each iteration process;
Figure QLYQS_83
Is a scalar with an initial value of +.>
Figure QLYQS_82
Gradually reducing in the iterative process;
Figure QLYQS_86
Is a control coefficient;
Figure QLYQS_67
And->
Figure QLYQS_72
Are all acceleration factors; />
Step4, repeating Step2 and Step3 until the algorithm converges;
step5. Output algorithm updates the location information.
CN202310436619.6A 2023-04-23 2023-04-23 User clustering and power distribution method of mMIMO-NOMA system based on meta-heuristic algorithm Active CN116155329B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310436619.6A CN116155329B (en) 2023-04-23 2023-04-23 User clustering and power distribution method of mMIMO-NOMA system based on meta-heuristic algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310436619.6A CN116155329B (en) 2023-04-23 2023-04-23 User clustering and power distribution method of mMIMO-NOMA system based on meta-heuristic algorithm

Publications (2)

Publication Number Publication Date
CN116155329A true CN116155329A (en) 2023-05-23
CN116155329B CN116155329B (en) 2023-07-28

Family

ID=86354722

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310436619.6A Active CN116155329B (en) 2023-04-23 2023-04-23 User clustering and power distribution method of mMIMO-NOMA system based on meta-heuristic algorithm

Country Status (1)

Country Link
CN (1) CN116155329B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117135737A (en) * 2023-10-24 2023-11-28 中国铁塔股份有限公司 Control method and device of base station power supply, electronic equipment and storage medium

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110225569A (en) * 2019-06-10 2019-09-10 桂林电子科技大学 A method of based on the WSNs clustering and multi-hop Routing Protocol for improving particle swarm algorithm
CN110418399A (en) * 2019-07-24 2019-11-05 东南大学 A NOMA-based Resource Allocation Method for Internet of Vehicles
CN113055088A (en) * 2021-03-09 2021-06-29 西安电子科技大学 Power distribution method of indoor visible light communication system based on improved particle swarm optimization
CN113329418A (en) * 2021-04-21 2021-08-31 中南大学 User clustering method, system and medium based on large-scale MIMO-NOMA system
CN113472409A (en) * 2021-05-19 2021-10-01 南京邮电大学 Hybrid precoding method based on PAST algorithm in millimeter wave large-scale MIMO system
CN114199248A (en) * 2021-11-30 2022-03-18 哈尔滨工程大学 An AUV Co-localization Method Based on Hybrid Metaheuristic Algorithm to Optimize ANFIS
CN115133969A (en) * 2022-07-01 2022-09-30 内蒙古大学 Performance Improvement Method of Millimeter-Wave Massive MIMO-NOMA System
CN115835358A (en) * 2022-11-24 2023-03-21 杭州电子科技大学 A Joint Optimization Method for User Grouping and Power Allocation in Uplink NOMA System
CN115833886A (en) * 2022-09-20 2023-03-21 内蒙古大学 Power control method of non-cellular large-scale MIMO system

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110225569A (en) * 2019-06-10 2019-09-10 桂林电子科技大学 A method of based on the WSNs clustering and multi-hop Routing Protocol for improving particle swarm algorithm
CN110418399A (en) * 2019-07-24 2019-11-05 东南大学 A NOMA-based Resource Allocation Method for Internet of Vehicles
CN113055088A (en) * 2021-03-09 2021-06-29 西安电子科技大学 Power distribution method of indoor visible light communication system based on improved particle swarm optimization
CN113329418A (en) * 2021-04-21 2021-08-31 中南大学 User clustering method, system and medium based on large-scale MIMO-NOMA system
CN113472409A (en) * 2021-05-19 2021-10-01 南京邮电大学 Hybrid precoding method based on PAST algorithm in millimeter wave large-scale MIMO system
CN114199248A (en) * 2021-11-30 2022-03-18 哈尔滨工程大学 An AUV Co-localization Method Based on Hybrid Metaheuristic Algorithm to Optimize ANFIS
CN115133969A (en) * 2022-07-01 2022-09-30 内蒙古大学 Performance Improvement Method of Millimeter-Wave Massive MIMO-NOMA System
CN115833886A (en) * 2022-09-20 2023-03-21 内蒙古大学 Power control method of non-cellular large-scale MIMO system
CN115835358A (en) * 2022-11-24 2023-03-21 杭州电子科技大学 A Joint Optimization Method for User Grouping and Power Allocation in Uplink NOMA System

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
LI WANGWANG; HUANG CHENGCHENG; XIE HUIXING; BI YAFAN: ""Study on the Release Property of Nitrogen in Sustained-release Fertilizer with Carrier of Bentonite"", 《IEEE》 *
金秀章，史德金，乔鹏: ""基于沙地猫群优化-最小二乘支持向量机的动态NOx排放预测"", 《中国电机工程学报》 *
高佰宏，刘朝晖,刘华: ""基于SCSO-GRU模型的网络流量预测"", 《计算机与现代化》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117135737A (en) * 2023-10-24 2023-11-28 中国铁塔股份有限公司 Control method and device of base station power supply, electronic equipment and storage medium
CN117135737B (en) * 2023-10-24 2024-01-26 中国铁塔股份有限公司 Control method and device of base station power supply, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN116155329B (en) 2023-07-28

Similar Documents

Publication Publication Date Title
CN111901862B (en) A method, device and medium for user clustering and power allocation based on deep Q network
CN101867402B (en) A MIMO system with adaptive antenna selection and its application method
CN114422363B (en) Capacity optimization method and device for unmanned aerial vehicle-mounted RIS auxiliary communication system
CN111294096A (en) A Channel Capacity Optimization Method of Smart Reflector MISO Wireless Communication System
CN104270820B (en) United vertical beam control and power distribution method in the extensive mimo systems of 3D
CN110011712B (en) Millimeter wave large-scale multi-input multi-output-oriented hybrid precoding method
Souto et al. A novel efficient initial access method for 5G millimeter wave communications using genetic algorithm
CN112019249A (en) Intelligent reflecting surface regulation and control method and device based on deep reinforcement learning
CN113329418B (en) User clustering method, system and medium based on large-scale MIMO-NOMA system
CN114040415A (en) DQN-DDPG-based resource allocation method assisted by intelligent reflector
CN116155329B (en) User clustering and power distribution method of mMIMO-NOMA system based on meta-heuristic algorithm
Yang et al. Deep reinforcement learning for RIS-aided non-orthogonal multiple access downlink networks
CN118282442B (en) Intelligent reflector-assisted large-scale MIMO resource allocation method and system
CN113543271B (en) Effective capacity-oriented resource allocation method and system
CN110536321A (en) The optimization method of energy efficiency and spectrum efficiency is weighed in 5G IOT communication
CN114285443B (en) Near-field broadband uplink MIMO transmission method assisted by dynamic super-surface antenna
CN112996118B (en) NOMA downlink user pairing method and storage medium
CN113766492A (en) Safe Massive MIMO network resource allocation method
CN119171923A (en) Interference suppression method of metasurface array, electronic device and communication system
CN115133969B (en) Performance Improvement Method of Millimeter Wave Massive MIMO-NOMA System
CN112398513A (en) Beam forming method of massive MIMO system
CN114745032B (en) Honeycomb-free large-scale MIMO intelligent distributed beam selection method
CN116470938A (en) Combined beam forming optimization method for IRS auxiliary communication service quality fairness
CN112836895B (en) Cell user division method based on improved support vector machine algorithm
Zhang et al. Joint optimization of energy efficiency and spectrum efficiency of single-station multi-satellite MIMO uplink system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant